Bleeding esophageal varices (BEV) due to cirrhosis is a common and highly lethal condition that is characterized by recurrent hemorrhage and early death in the majority of patients. Many types of therapy have been used for BEV, but it is not known which one is best and none has been shown unequivocally to improve long-term survival. The objective of the proposed study is to compare two widely used therapeutic modalities, emergency portacaval shunt (EPCS) and emergency and longterm endoscopic sclerotherapy (EST), with regard to (a) mortality rate, (b) duration of life, (c) quality of life, (d) direct and indirect economic costs, and (e) treatment failure, defined precisely in terms of number of blood transfusions required because of recurrent BEV. Unselected patients ("all comers") with BEV due to cirrhosis will be randomized to either EPCS or EST. The diagnostic workup will be completed rapidly, EPCS or EST will be done within 8 hours of initial contact, and lifelong follow-up will be conducted in a special clinic. Treatment crossover (EPCS to EST, and EST to PCS) will be done only after primary treatment failure has been declared. The study will be monitored by an external Advisory Committee of senior scientists, and the data will be processed and analyzed by an external biostatistical firm with widely-recognized expertise in clinical trials. The study is unique in that it will be a community-vide endeavor in which 65 San Diego area physicians and two governmental hospitals will promptly refer patients with BEV to UCSD Medical Center for inclusion in the study. By clearly determining the most effective treatment of BEV, the results of this study should have substantial impact on the survival and quality of life of the bleeding cirrhotic population, and should provide important information on the economic costs and benefits of these two forms of therapy.